Mixed CNS glial cell cultures were obtained from neonatal rat brain using a modification of the so-called "shake-off" technique 2526 as we described previously 27. Following shakeoff of cells from the astroglial bed layer, the time for partial removal of microglia by adherence to plastic was 1 hour prior to plating on poly-lysine coated flasks. Cells were maintained in defined medium containing 2% fetal bovine serum for 6–8 days, then treated with the cytokines. The composition of cultures was examined by indirect immunofluorescence (IF) with antibodies to phenotypic markers for different cells types: glial acidic fibrillary protein (GFAP) for astrocytes 28 (Chemicon, Temecula, CA); galactolipids (GalL) for oligodendrocytes 2829; A2B5 for oligodendrocyte precursors 30 (ATCC, Bethesda, MD); ED-1 for microglia 31 (Serotec, Raleigh, NC), Thy1.1 for fibroblasts 32 and in glial cultures some astrocytes 33; anti-neurofilament heavy chain (NFh) for neurons 34 and anti-factor VIII for endothelial cells (Dako Corporation, Carpinteria, CA).

The Th1 cytokine mixture included the rat recombinant cytokines interleukin-2 (IL-2), interferon-γ (IFN-γ) (R&D Systems, Inc, Minneapolis), tumor necrosis factor-α (TNF-α; BD PharMigen, San Diego, CA) and mouse granulocyte-colony stimulating factor (G-CSF; PeproTech, Rocky Hill, NJ).

The M/M cytokine mixture included the rat recombinant cytokines IL-1α and IL-1β, IL-6, IL-12p40 (all from R&D Systems, Inc) and TNF-α. These cytokines would be considered proinflammatory products of M1 macrophages or microglia 35.

The Th2 cytokine mixture included the rat recombinant cytokines IL-4, IL-5, and IL-10 (all from R&D Systems, Inc), mouse G-CSF and purified porcine transforming growth factor-β1 (TGF-β1; R&D Systems, Inc.). In the cognate immune system, in some species, TGF-β1 is considered by some to be the product of so-called Th3 cells. TGF-β1 is also important in the development of another population of T-cells called regulatory T-cells (Treg cells) which are phenotypically characterized as CD4+/CD25 high+/Fox3 3637. These Treg cells may also secrete TGF-β1.

Cytokine mixtures contained 10 ng/ml of each of the constituent cytokines as is typically employed many in vitro studies of cytokine biology. For each experiment, four groups of three T75 flasks per group were incubated either with mixtures of Th1, Th2, M/M cytokines or additional medium (control) for 6 hours. Three sets of separate experiments consisting of control, Th1, M/M and Th2 stimulated cultures were performed.

As reported 1516, we examined the cytokine-induced effect on cell death in mixed CNS glial cell cultures by incubating cultures from 6 hours to 4 days with the cytokine mixtures. Cell death was determined by uptake of 0.4% trypan blue 38.

Cultures were washed and frozen after 6 hours of incubation with cytokine mixtures or additional medium. RNA was extracted employing TRIzol (Gibco BRL, Grand Island, NY) followed by Qiagen RNeasy kits (Qiagen, Valencia, CA). The RNA was quantitated at A_{260 nm} and the quality was assessed by at A_{260 nm}/A_{280 nm}. The 28S/18S ratio was assessed using a Bioanalyzer 2100 (Agilent Technologies, Wilmington, DE), and was > 1.7 for all samples.

Biotin-labeled RNA fragments were prepared and hybridized to the Affymetrix rat RG-U34A microarray at 45°C for 16 hours, as previously described 1516. Subsequent signal amplification was performed employing biotinylated anti-streptavidin antibody. The RG-U34A chip contains 7,985 genes. The control and three cytokine-incubated cultures from one experiment were analyzed with one gene chip for each sample and three separate experiments using different cultures were analyzed.

Data were analyzed by comparing the average of the replicates from each of the separate 3 sets of experiments. Affymetrix data were analyzed with dChip v1.2 to correct for background and calculate gene expression values 39. We analyzed values from 3 separate experiments employing the t-test in GeneSpring comparing Th1, M/M and Th2 with control. Multiple testing analyses that compare all 7,985 genes at different levels of stringency using the Bonferoni and false discovery value (FDV) are statistically most rigorous, but at such high levels of stringency, there were very few changes that reached statistical significance. In order to increase sensitivity and allow identification of potentially important biologic changes, we employed a lower level of stringency 1516. In these screening studies at a single time point, we have arbitrarily chosen to represent as probably significant those genes in which the mean expression was > 2 fold (upregulation) or < -2 fold (downregulation) compared to expression in controls (p < 0.2) 15. We believe this is reasonable given that our experiments consisted of biological replicates that are prone to greater variability than experimental replicates. A similar p value was used in a gene array analysis of MS lesions 2. The recent literature suggests that a 2-fold cut-off using the Affymetrix platform produces a low false positive rate 40.

Expression of message for iNOS was analyzed by QRT-PCR on an ABI 7500 Fast System, using ABI Taqman rat specific gene expression assays. RNA was extracted as above and reversed transcribed. Relative expression levels were calculated with GAPDH as the internal reference, using the delta-delta Ct method 41. The values from the treated cultures were compared to those from control. Those ratios were averaged for the three experiments, then expressed as fold changes in the treated cultures relative to control for comparison with the gene array results. Each PCR value represents the average from 2–3 separate experiments.

As in our earlier papers, cultures consisted of approximately 35% each of oligodendrocytes and astrocytes and 10% microglia. The remaining cells were glial cell precursors including A2B5 positive oligodendrocyte precursors. Endothelial cells and neurons were not present. Viability was > 98% in all cultures control and cytokine stimulated, at all time points examined (6 hours to 4 days) demonstrating the lack of cytotoxicity under these conditions.

In the preceding papers we first described changes in CNS glia in genes for proteins predominantly associated with the immune system including major histocompatibility molecules, several adhesion and extracellular matrix molecules, cytokines and chemokines and their receptors and complement components 15. Because of our interests in the effects of cytokines on the production of factors important in oligodendrocyte, axonal and neuronal function, in a second paper we compared the effects of the different cytokine mixtures on expression of genes for neurotrophins, growth factors, related receptors and structural proteins 16. This third paper summarizes our findings for cytokine-induced changes in glial expression of genes for proteins associated with metabolism, signaling and regulation as well as neurotransmitters and ion channels. As noted, this is a series of screening experiments and therefore Tables 1 and 2 were prepared using the criteria of > 2 fold (increased expression) or < -2 fold (decreased expression) with a p value of < 0.2 for one or two replicates of the gene transcript 1516. Unknown genes (ESTs) are not presented.

All three cytokine mixtures had regulatory effects on message levels for a wide range of message levels for neurotransmitters and their receptors as well as on transporters involved with transmitters including glutamate, adrenergic, cholinergic, glycine, serotonergic, dopaminergic and purinergic systems (Table 1). The only adrenergic receptor affected was alpha 2 c-4, upregulated 2.5 fold (p < 0.05) by Th1 cytokines. Among cholinergic receptors, the largest change was for nicotinic cholinergic receptor alpha5, downregulated -2.3 fold (p < 0.05) by Th1. Dopaminergic receptors A3 and D1 were markedly downregulated -8 to -14 fold by Th1 and Th2 cytokines. Among several changes in glutamate receptors, Th1 upregulated ionotropic glutamate receptor delta 1 by 2.7 fold (p < 0.01), but markedly downregulated metabotropic glutamate receptor 7b by -9.5 fold (p < 0.01). Neuropeptide Y receptor 5 was downregulated by both Th1 and MM cytokine, -18 fold (p < 0.10) and -8 fold (p < 0.20), respectively, while the substance P precursor preprotachykinin A was downregulated -7 fold (p < 0.10) by Th2. For purinergic receptors, the most robust changes were -3 fold (p < 0.05) downregulation of P2X1by Th2, and upregulation of P2Y2 by MM and Th2, 3.5 fold and 2.4 fold, respectively, both p < 0.05.

Th1, M/M and Th2 cytokines had primarily downregulatory effects on expression of a very large number of genes for proteins that are components of ion channels including Na, K, Ca and Cl channels, both voltage-gated and non-voltage gated (Table 1). For example, Th1 and Th2 downregulated the voltage-gated alpha 1D L type Ca++ channel by -4 and -7 fold respectively, both p < 0.05. A large number of K+ channels were downregulated by Th2 cytokines, with fewer downregulated by Th1 or MM cytokines. The voltage-gated 1 alpha sodium channel was robustly downregulated by Th1 and MM cytokines, -9 fold (p < 0.05) and -7 fold (p < 0.10) respectively, while Th2 cytokines uniquely downregulated the 1 beta isoform, -2.5 fold, p < 0.01.

In addition to the effects on ion channels shown in Table 1, there were effects on several ATPase ion exchangers. With the exception of upregulation of Ca++ATPase (plasma membrane 1) by M/M cytokines, several ATPase ion exchangers were downregulated by each of the cytokine mixtures (Table 1).

The cytokine mixtures induced up and down regulation of several genes for proteins involved in control of apoptosis (Table 1) including caspase 2, downregulated -3 fold, p < 0.05 by both MM and Th2 cytokines, and caspase 7, downregulated -3 fold by Th1 and MM cytokines.

There were cytokine mixture-induced changes in expression in genes of some mitochondrial proteins which are listed in Table 1, including a 4–6 fold (p < 0.05) upregulation of super oxide dismutase 2 (SOD2) by Th1 and MM, and an apparent -4 fold (p < 0.01) decrease for 16S mitochondrial ribosomal RNA.

The genes for several proteins involved in glutathione metabolism and secretion were regulated by the different types of mixtures of cytokines (Table 1). Subunits of glutathione S-transferase were generally downregulated, while both Th1 and MM upregulated P-glycoprotein(multi-drug resistance protein) by 5 fold, p < 0.10 and p < 0.05, respectively.

Th1 cytokines markedly upregulated junB, CREB and the p105 subunit of NF κB. Both Th1 and M/M cytokines altered expression of genes for several other transcription factors, while the Th2 cytokines had minimal effects (Table 2). All three cytokines markedly upregulated message levels of junB (2–3 fold, p < 0.01) and downregulated hepatic nuclear factor alpha (-3 to -7 fold, p < 0.05.). Both Th1 and M/M cytokines upregulated the expression of the gene for NF kappa B p105 subunit and downregulated the aryl hydrocarbon receptor (Table 2).

Th1 and M/M cytokines upregulated the gene for peroxisome proliferator activator receptor δ (PPAR δ) by 2 fold (p < 0.10), and downregulated the gene for PPAR γ by -5 fold (p < 0.01). Th2 cytokines downregulated the gene for PPAR γ but had no effect on PPAR δ (Table 2).

The cytokine mixtures had effects on the expression of genes for many signal transduction molecules, some of which are presented in Table 2. Among these are STAT1, 3 and 5, JAK2, homer and protein kinase 2 B. Of interest, only one gene in this category was affected by Th2 cytokines, stress activated protein kinase alpha 2, upregulated 3 fold (p < 0.05).

Th1, M/M and Th2 cytokines had effects on the genes for several proteins involved in synthesis, degradation and intracellular transport of proteins including synucleins, proteasome subunits, ubiquitin conjugating enzymes and heat shock protein (HSP) 70 kDa (Table 2). For example, proteasome (macropain) alpha 6 was upregulated 10 fold (p < -.01) by Th1 cytokines, while the ubiquitin-conjugating enzyme E2D 2 was downregulated -4 to -6 fold (p < 0.05) by all three cytokine mixtures.

As noted in Table 2, the genes for several proteins involved in lipid metabolism and signaling were regulated by the different cytokine mixtures. For example, fatty acid CoA ligase long chain 4 was downregulated -4 to -5 fold (p < 0.05 fold) by Th1 and Mm cytokines, while UDP-glucose ceramide glycosyl transferase was upregulated 3 fold (p < 0.05). With regard to lipid signaling, CDP diacylglycerol synthase was markedly upregulated 6–10 fold (p < 0.01) by all three cytokines, while EDG (endothelial sphingolipid GPCR) was downregulated -6 fold (p < 0.01) by MM cytokines.

Specific genes regulating proteins involved in sterol and vitamin D metabolism were also regulated by the cytokine mixtures (Table 2). Thus, 17-beta hydroxyl sterol reductase was downregulated -4 to -5 fold (p < 0.05) by Th1 and Th2 cytokines, and testosterone 6-beta hydroxylase was downregulated -7 fold (p < 0.05) by MM cytokines.

There were a large number of genes for proteins not included in the above categories that are potentially of importance in understanding the pathogenesis of MS, as well as protective and reparative mechanisms. These are listed in Table 2. Of note, Th1 and MM cytokines upegulated nitric oxide synthase 2 (iNOS) by 15 fold (p < 0.05), while Th2 cytokines markedly downregulated genes for angiotensin receptor type 1b (-5 fold), beta tubulin 2b (-12 fold), and chapsyn-110 (-29 fold), all p < 0.05, but upregulated arginase and low density lipoprotein receptor 1 by 5 fold (p < 0.05).

We validated expression changes in three genes by QRT-PCR: iNOS, the enzyme that synthesizes NO from arginine; arginase, the enzyme that breaks down arginine, thus limiting production of NO; and P-glycoprotein (multi-drug resistance 1), an ABC transporter involved in regulation of glutathione levels. As noted in Table 3, we were able to confirm striking upregulation of the gene for iNOS by Th1 and M/M at 6 hours employing QRT-PCR. Although no effect on the gene for iNOS expression was observed at 6 hours in response to Th2 cytokines on microarray, we detected modest downregulation employing QRT-PCR. For arginase, we confirmed upregulation by Th2, with no change induced by Th1; however, in contrast to the array results, PCR indicated some upregulation of arginase by MM at 6 hours, rather than no change. For P glycoprotein, PCR showed upregulation by Th1 (in two of three analyses) and MM, as on the gene array, but also indicated a modest increase with Th2 rather than no change. The results for these three genes show relatively good agreement, and indicate that the arrays are not giving false positive results, but in some instances may give false negative results, suggesting that PCR may be more sensitive than the gene array.

Glial cells have been reported to express different neurotransmitters and receptors as well as transporters for these transmitters 2342434445. With increasing evidence that both oligodendrocyte and neuronal/axonal damage may be caused by glutamate induced toxicity 4647484950, and that other glutamate receptors may be protective 51, the effects of cytokine mixtures on different GluR may influence and modify the effects of glutamate. AMPA, kainate and NMDA receptors may be important in oligodendrocyte toxicity in MS and EAE 525354555657 whereas upregulation of metabotropic GluRs (mGluRs) may provide protection 58. The effects of upregulation of AMPA, NMDA and kainate GluR on neuronal death are well recognized 59. It is of interest that Th1 and M/M upregulated GluRs associated with cell toxicity whereas all three cytokine mixtures markedly downregulated metabotropic mGluR7b. The group III mGluRs, including mGluR7, inhibit production of RANTES induced in astroglia by TNF-α or IFN-γ 60. We previously reported upregulation of the gene for RANTES by Th1 and M/M cytokines 15.

The role of other transmitters in glial cell function is not as well understood. In addition to receptors for well established neurotransmitters and classically described ion channels (Ca++, K+ and Na+), we detected effects on genes for the purinergic P2X receptors, some of which serve as ligand gated ion channels 61, and the P2Y receptors which act as G protein coupled receptors when ligated by extracellularly released nucleotides, as occurs in inflammation and other stressful conditions within the CNS. The purinergic receptors modify the response of astrocytes to cytokines such as IL-1β and TNF-α and modify astrocyte functions including Ca++ influx 626364, as well as modulating transport of other ions 65. The P2X7 receptor is found in resting and activated microglia in epileptic brain and several other neurologic diseases 66, and plays a role in microglial proliferation 61 and migration 23. The P2X7 receptor is expressed in reactive astrocytes and microglia/macrophages in MS lesions 62, and is transiently upregulated by the M/M cytokine IL-1β in cultured fetal human astrocytes, resulting in increased iNOS activity 64. We noted modest downregulation of the gene for the P2X7 receptor with M/M at 6 hours in our mixed glial cultures, suggesting that the mixture of cytokines or the presence of other glial cell types may modulate the glial responses to IL-1β, or that upregulation seen in vivo and in vitro in other studies may be a secondary response occurring at later time points.

The different cytokine mixtures had variable effects on a large number of receptors for several other transmitters including dopaminergic, serotonergic, cholinergic, adrenergic and melanocortin as well as the previously discussed purinergic receptors. The downregulation of the genes for the D1 and D3 dopamine receptors by the cytokine mixtures was especially striking. All three types of glia are known to express dopamine receptors, and D3 dopamine receptors may play roles in oligodendroglial differentiation and myelination 67, as well as protection of oligodendrocytes against glutamate oxidative stress and oxygen glucose deprivation 68. Binding of neurotransmitters to their receptors on microglial cells seem to be important in microglial function 23.

It is also of interest that several of the neuronal nicotinic acetylcholine receptors are involved in downregulation of proinflammatory immune reactions in the systemic immune system, in particular the nicotinic α7 receptor 6970, which we found downregulated by both M/M and Th2 cytokine mixtures. Further, attenuation of cholinergic signaling with the acetylcholinesterase inhibitor physostigmine results in inhibition of CNS inflammation and development of EAE 71. Conversely, we found that M/M cytokines upregulated the gene for the acetylcholinesterase T subunit, which could lead to increased acetylcholinesterase and a decreased "protective" cholinergic response.

Expression of genes for several transporters for glutamate and glycine were also observed along with changes in the genes for the R5 receptor for neuropeptide Y receptor 5 and preprotachykinin A (precursor of substance P). The role of such receptors and transporters in glial cells is not clear. Of interest, changes were found in activity of neurotransmitter-induced early genes 9, 10 and 12. It is not known if these cytokine mixtures have a similar effect on the same genes and their proteins in various subpopulations of neurons. If they do, this would have major implications on neuronal and axonal dysfunction in MS and other diseases characterized by CNS inflammation and/or microglial activation 72, as well as symptoms in patients with MS such as depression, memory loss, abnormalities in other cognitive functions, fatigue and pain.

We observed cytokine induced changes in gene expression for many ion channels including Na+, K+ and Ca++ channels. It is well established that glial cells have a wide variety of ion channels which are important in glial cell function 217374757677 and that expression of some of these have been reported to be affected by cytokines in glial cells and neurons, as well as other types of cells 2178798081. Cytokine effects on ion channels and ion exchangers clearly are important in axonal and neuronal function and viability as well as likely contributing to symptomatology in patients with MS 8283848586. Changes in genes for ion channels have been reported in the CNS in MS and EAE 287.

There have been reports of inflammatory mediators such as inducible nitric oxide synthase (iNOS) inducing upregulation of certain Na channels in neurons 8488. We observed significant effects on gene expression for a wide variety of ion channels in glial cells at 6 hours of incubation suggesting a direct effect of cytokines on expression of genes for some or all of these channels in glia. Changes in the distribution of ion channels could contribute to glial cell dysfunction. If similar changes were induced directly in neurons/axons, these changes could contribute to plasticity as well as to axonal and neuronal cell death. While such changes in neuronal ion channels and failure and reversal of ion exchangers, particularly Ca++ exchangers 8990, could result from failure of mitochondrial energy metabolism 1091, our results also raise the possibility of axonal dysfunction and ultimately death by direct effect of cytokines on expression of genes for ion channels and ATPase ion exchangers (see below). The cytokine mixtures also likely regulate ion channels on inflammatory cells such as lymphocytes, and ion channels are known to affect lymphocyte function 92.

Th1 downregulated the genes for Na+/K+ ATPase, α4 polypeptide and Ca++ transporting ATPase. M/M upregulated the genes for Ca++ transporting ATPase and Ca++ ATPase, plasma membrane 1. Th2 likewise downregulated the genes for several ATPase ion transporters including Na⁺/K⁺ ATPase α4 polypeptide, Ca++ transporting ATPase, Ca++-pumping ATPase isoform 4, H+/K+-ATPase α2 gene, alternatively spliced and H⁺/K⁺ ATPase, nongastric, α polypeptide, nongastric; and the Na⁺/H⁺ ion exchanger (Table 1).

The possible role of oligodendrocyte cell death through apoptosis via caspases 93 or via other pathways to cell death 94 in MS lesions continues to be controversial, and it is likely that apoptosis as a mediator of oligodendrocyte death varies in different lesions 95. Neuronal cell death by what appears to be apoptosis is also seen 96. Up and downregulation of expression of various genes for proteins involved in apoptosis including caspases and Bcl X were noted. Th1 and M/M cytokines both induced upregulation of the genes for caspase 7, a downstream effector caspase involved in caspase-dependent apoptosis 97, and Bcl X, a protein which inhibits apoptosis 9899. M/M cytokines downregulated the gene for caspase 2, a caspase implicated in oligodendrocyte cell death via the p75 receptor 93. The gene for cytolysin (a constituent of lymphocyte toxic granules) was downregulated by both Th1 and Th2 cytokines, and Th2 and M/M cytokine mixtures both downregulate the gene for the protein programmed cell death 2.

Mutations in the gene for the protein huntingtin (Htt) result in Huntingon's chorea. Htt interacts with several proteins. One of these proteins is called htt-interacting protein (HIP-1). When HIP1 is bound to normal Htt, it inhibits apoptosis in certain neurons and Htt seems to be involved in endocytosis as well 100101102103. In addition abnormal huntingtin interferes with normal ubiquitin-proteosome function and one could readily postulate that downregulation of proteins such as HIP-1 that interact with htt could also lead to abnormal protein aggregation such is seen in many degenerative diseases including Huntington's disease, where it is the htt that is qualitatively abnormal 104. There was downregulation of the gene for HIP-1 by Th1 and M/M cytokines.

Changes in expression of mitochondrial protein genes, including genes associated with some apoptotic pathways, were noted (see below).

Changes in mitochondrial related genes have been noted in MS cortical gray matter in patients with long-standing chronic MS 10, and failure in mitochondrial associated energy metabolism may be important in axonal and neuronal degeneration and cell death 8991. Most of the detected changes were reduced expression of genes, particularly components of complex I, III and IV. Decreased expression of COX subunits I and IV (Complex IV) has been detected in oligodendroglia, astroglia and axons, but not in microglia, in acute Type III MS lesions 12. In our CNS glial cultures, we found predominately downregulation of genes associated with mitochondria. Th1 cytokines upregulated genes for hexokinase II and superoxide dismutase 2 (Mn++ SOD2), and downregulated genes for NADH dehydrogenase (Complex I), COX VIa and COX VIc (Complex IV), ferredoxin 1, and perhaps for 16s ribosomal RNA, which is a component of the large subunit of the mammalian mitochondrial ribosome, responsible for synthesis of 13 proteins localized in the inner mitochondrial membrane 105. This latter finding has yet to be confirmed. M/M cytokines upregulated genes for SOD2 and down regulated the genes for NADH dehydrogenase and creatine kinases. Th2 cytokines down regulated genes for NADH dehydrogenase and 16S ribosomal RNA. Th1 down regulated the gene for SOD3, an extracellular Cu⁺⁺/Zn⁺⁺ SOD. None of the cytokine mixtures had an effect on the gene for the mitochondrial protein Cu⁺⁺/Zn⁺⁺ SOD1; some familial forms of amyotrophic lateral sclerosis (ALS) are associated with mutations in this gene 106. One could postulate that inflammatory cytokines, perhaps products of activated microglia, at first stimulate transcription of genes for some mitochondrial enzymes, but decreased expression of the 16S subunit of the mitochondrial ribosome could lead to ongoing downregulation of genes and their proteins critical for mitochondrial function in oligodendrocytes and neurons.

We observed effects on the genes for several proteins involved in glutathione metabolism and secretion. Glutathione serves a protective function by reducing the effect of free radicals produced via oxidative stress 107108, and the cytokine mixtures had significant effects on genes for several proteins involved in synthesis, regulation and release of glutathione. Th1 upregulated the gene for P-glycoprotein/multidrug resistance protein 1/MDR1 (P-glycoprotein/ABCB1), as did M/M cytokines. MDR1, which in addition to inhibiting the therapeutic effects of drugs, has effects on vascular structures and influences secretion of glutathione by cells such as astrocytes 109110111. In addition to astrocytes, it has been detected in microglia, oligodendrocytes, endothelial cells and neurons 112113. Glutathione is more abundant in astrocytes than in other brain cell types, which may contribute to the relative resistance of astrocytes to ischemia and other pathologic processes that involve oxidative stress. Changes in glutathione may also contribute to the relative vulnerability of oligodendrocytes and precursors at different stages of maturation to oxidative stress 107114115116. Glutathione also modulates prostaglandin metabolism 117. We describe cytokine modulation of expression of several genes associated with prostaglandin metabolism 15, and prostglandin D synthase has been reported to be upregulated in MS lesions 2.

Some genes for transcription factors were upregulated, quite predictably, such as NF-κB p105 in the presence of IL-1 or TNF-α (Th1 and M/M cytokines). Cyclic AMP response element binding protein 1 (CREB) was upregulated in response to Th1 cytokines.

The upregulation of the gene for CREB by Th1 cytokines may result from the effect of TNF-α118, although M/M cytokines which also contain this cytokine did not appear to have the same effect.

Jagged 1 is a transcription factor reported to be upregulated in astrocytes by TGF-β and through Hes and notch1 leads to inhibition of myelination 119, although its presence may not be necessary to inhibit myelination 120. Given the report of upregulation of jagged 1 by TGF-β in astrocytes 119, an unexpected finding in our experiments was upregulation of jagged 1 by Th1 and M/M cytokine mixtures which do not contain TGF-β and the failure of the Th2 cytokine mixture, which contains this cytokine/growth factor, to upregulate jagged 1. These differences could relate to differences in the target tissues (single cell types versus mixtures of different cell types), species and/or effects of a single cytokine versus mixtures of cytokines, effects of some of the induced proteins and their influence on downstream signaling. At 6 hours none of the cytokine mixtures had any discernable effect on expression of notch1 or Hes.

In addition we observed effects on gene expression of other transcription factors including hepatic nuclear factors (HNF), POU and elongation initiation factor 5, important in initial stress responses.

PPARs are nuclear receptors originally associated with lipid metabolism but subsequently found to also be involved in cellular differentiation. Th1 and M/M cytokines both upregulated the gene for PPAR δ and down regulated the gene for PPAR γ, whereas Th2 cytokines down regulated the gene for PPAR γ and had no effect on the gene for PPAR δ. TNF-α, a component of both the Th1 and M/M cytokine mixtures, has been reported to down regulate the gene for PPAR δ 121. This is another potential example of differences in the effects of single cytokines versus mixtures of cytokines. Ligation of PPAR γ results in down regulation of inflammatory responses and can inhibit EAE 122123 and has lead to studies to evaluate such agents, which are used in the treatment of diabetes and hyperlipidemia, as therapy for MS. Activation of PPAR δ with different ligands than those that react with PPAR γ causes activation and accelerates differentiation of oligodendrocytes in vitro 124. How the differential regulation of the PPARs affects inflammation and the potential for favorably influencing remyelination through these receptors is not as yet clear.

Th1 and M/M cytokines markedly upregulated Janus kinase 2 (JAK) as well as several members of the STAT family, in keeping with the well known activation of the JAK/STAT pathway by inflammatory cytokines. Studies in brain ischemia indicate that increased signaling via the JAK/STAT pathway occurs predominantly in microglia rather than astroglia or neurons 125. We found a 10 fold increase in STAT 1 with Th1 cytokines, consistent with an increase in STAT 1 identified by gene array analysis in both chronic and active MS lesions 126.

The gene for homer, a key protein in Group I metabotropic GluR signaling, was upregulated by Th1 and Th2 cytokines 127128129130. In addition there were effects on expression for genes of protein kinase C, protein kinase A, protein tyrosine kinase 2B, CaM kinase II, Rho family GTPase, receptor serine threonine kinase, lyn protein non-receptor kinase and fyn-related kinase. Fyn, the only src family kinase, is upregulated during oligodendrocyte differentiation 131 and signals through Rho family GTPases to regulate their morphologic maturation 132.

Th1 cytokines upregulated the genes for α synuclein and for several proteasome proteins and ubiquitin-like protein (NEDD 6), and down regulated the genes for proteasome subunit R-RING (although as noted upegulated other transcripts of the R-RING subunit), β synuclein (SYN 2), receptor serine threonine kinases, ubiquitin ligase (NEDD 4), and ubiquitin conjugating enzyme E2D 2. M/M cytokines upregulated the genes for NEDD 6 and SYN2, and downregulated the genes for ubiquitin-conjugating enzyme E2D 2, NEDD 4 and proteasome subunit R-RING. Th2 cytokines downregulated the genes for ubiquitin conjugating enzyme E2D 2 and similar to ubiquitin-conjugating enzyme E2D 2. If changes in the expression of these genes result in changes in the level of these proteins, it would imply that inflammation could contribute to the changes in these proteins seen in sporadic forms of several degenerative disorders where synucleins and ubiquitin aggregation have been described 133134. The synucleins, considered neuronal proteins, are involved in synaptic function and have chaperone functions as well 135136. Th1 cytokines upregulated α-synuclein, which has been detected transiently in rat oligodendrocytes in vitro 137 and in inclusions in glial cells in some CNS diseases including multisystem atrophy (MSA) 138. Proteasomes are involved in transport of protein degradation products as well as in transport of MHC proteins and antigen within antigen presenting cells (APC).

Heat shock proteins (HSP) are upregulated in response to several types of cell stress stimuli 139. One of several functions of HSP is acting as chaperones to help in normal transport of other proteins within the cytoplasm of many cell types. M/M cytokines upregulated the gene for heat shock protein (HSP) 70 kDa. Interestingly upregulation of the gene for α/β crystalline which also serves as a stress response protein has been reported to be increased in MS lesions 2.

Th1 cytokines altered gene expression for several enzymes involved in synthesis of fatty acids and phospholipids (Table 2). Both Th1 and M/M cytokines downregulated message of the gene for HMGCoA reductase, the principal regulatory enzyme for cholesterol and other isoprenoids. Interestingly statins, which are inhibitors of this enzyme, are being tested as treatment for MS 139140141142 because they inhibit experimental autoimmune encephalomyelitis (EAE), an animal model for MS. The mechanisms include decreased farnesylation causing a Th1 to Th2 shift and monocyte/macrophage inflammation 141143144145, and perhaps alteration of other signaling pathways 16.

UDP-glucose:ceramide glycosyltransferase is upregulated in the presence of TNF-α (Th1 and M/M cytokines). This enzyme is involved in ceramide metabolism as part of both ceramide induced cell death via TNF-R type I signaling pathways, as well as catalyzing the initial step in ganglioside synthesis. Th1 cytokines also downregulated the gene for UDP-galactose ceramide galactosyltransferase (member 8 of the UDP-glucuronosyl transferase family). This enzyme is markedly upregulated during differentiation of oligodendroglia and synthesizes galactocerebroside, the major glycolipid of myelin and precursor to sulfatide. An early response to inflammatory cytokines has not been previously reported for the gene or the protein.

Notably Th1 cytokines upregulated the gene for phospholipid scramblase, which translocates phospholipids from one surface of the plasma membrane to the other.

In our initial paper we reported that Th1 and M/M cytokines induced robust upregulation of genes for ABC transporter 1, which among its several functions, translocates phosphatidyl choline and cholesterol to the outer membrane leaflet in astrocytes and neurons 146, and for ABC transporter 2, active in oligodendrocytes during myelination 147. The ABC transporters are also important in intraceullar transport of other proteins including peptide epitopes with MHC class I molecules 15.

Th1 cytokines downregulated the gene for diacyl glycerol kinase beta, which phosphorylates diacyglycerol to produce phosphatidic acid, leading to termination of diacylglyceryol signaling via PKC, Ras GTPase and other signaling pathways.

IL-2, one of the components of the Th1 mixture, upregulates diacylglycerol kinase in myelin 148, raising the possibility that oligodendroglia are upregulating this gene in response to the Th1 cytokines. In addition, the gene for CDP-diacylglycerol synthase, the enzyme that synthesizes phosphatidyl inositol from phosphatidic acid, is robustly upregulated by both Th1 and M/M cytokines, suggesting a switch from diacylglycerol to phosphatidyl inositol mediated signaling pathways. Conversely, the gene for myo-inositol monophosphatase was downregulated by M/M cytokines; this enzyme, the key enzyme inhibited by lithium, generates free inositol from inositol-3-phosphate derived from glucose-6-phosphate, and regulates levels of inositol available for synthesis of phosphatidyl inositol and its multi-phosphorylated derivatives critical for intracellular signaling and trafficking as well as calcium homeostasis. It is of note that lithium is currently being evaluated as a treatment of amyotrophic lateral sclerosis based on its effects on inositol pathways 149150151152.

Sphingosine-1-phosphate plays a key role in cell survival and inflammatory responses 153; the gene for one of its receptors, EDG (endothelial sphingolipid GPCR) was down regulated by both M/M and Th2 cytokines. There has been a phase II trial in patients with MS of an oral agent called FTY72, which binds to the EDG (S1P) receptor (endothelial differentiation sphingolipid G-protein coupled receptor) 154. A large Phase III study is underway. Inhibition of this receptor both blocks emigration from and favors homing of lymphocytes to secondary lymph structures, ostensibly without affecting T-cell viability or inhibiting memory T-cells. In experimental animals other inflammatory cells, such as monocytes and mature dendritic cells, are also affected and the protein is also found on endothelial cells. The drug has also been used in studies of treatment of other immune disorders 155156157158159160 The S1P receptor EDG is also found in the CNS on glial cells 161162163. The roles of S1P and its G-coupled receptor in the normal CNS are not known. It has recently been shown that activation of S1P results in changes in glial cells in vitro 164165. If FTY720 gains access to the CNS there is the potential to modulate the activity of S1P with uncertain consequences for the patient.

Several genes coding for enzymes involved in steroid metabolism were downregulated by each of the three cytokine mixtures, including the gene for testosterone 6-beta-hydroxylase, markedly downregulated by M/M cytokines. Of note, Th1 cytokines upregulated the gene for vitamin D3 25-hydroxylase, the enzyme catalyzing the first step in activation of dehydrocholesterol to the active hormone, 1, 25-hydroxy vitamin D3. However, both M/M and Th2 cytokines down regulated 25-OH vitamin D3 24-hydroxylase, a key step in the inactivation of the active form of vitamin D3 166. Both are mitochondrial enzymes and members of cytochrome p450 family. In several studies vitamin D3 dietary supplementation prevented the onset and progression of EAE. In MBP-induced EAE in mice, the treated animals showed marked decreases in chemokines, iNOS and CD11b+ recruitment into the CNS, perhaps due to activated T cell apoptosis 167. One large study found that vitamin D3 supplementation reduced the risk of developing MS 168, while four smaller studies suggested a reduction in exacerbations (reviewed in Brown, 2006) 169. Our findings suggest that both M/M and Th2 cytokines might act to attenuate the effects of the active forms of vitamin D3.

The classically proinflammatory Th1 and M/M cytokines markedly upregulated the gene for iNOS, a critical protein in generation of NO, which gives rise to related reactive oxygen species such as peroxynitrite 170. Increases in iNOS have been reported in the CNS in EAE and in MS 171172173174. There is evidence that NO could directly or indirectly, by forming peroxynitrite, damage oligodendrocytes, myelin and neurons/axons 175. Reactive nitrogen species can also influence neuronal Na channels and thus cause damage, especially with rapid firing bare axons 171176. It has also been suggested that NO could have an immumodulatory effect on inflammatory cells. NO production in inflammatory cells and in glial cells is induced by iNOS. As described in Results, employing QRT-PCR we confirmed the upregulation of expression of the gene for iNOS by Th1 and M/M cytokine mixtures and also found modest downregulation of the gene in response to Th2 cytokines (Table 3).

Galanin is a peptide in the CNS and PNS which is upregulated in response to injury 177178. While originally described in various neurons it has been demonstrated in glia as well.179180 and has a positive effect on neurite growth, cell survival and regeneration 181182183 as well as involved in interactions with hormones 184185, pain signaling pathways 186187 and other CNS functions 188189. Galanin receptors are also co-localized with cholinergic receptors in astroglia 190. During oligodendrocyte differentiation, the gene for galanin is markedly downregulated 191. Therefore upregulation by Th1 and M/M cytokines may represent an early attempt of oligodendroglia to return to a less differentiated state, one capable of proliferation.

The alpha-fetoprotein that is increased in the serum of women in the last trimester of pregnancy has been shown to have immunosuppressive effects in EAE as well as in experimental autoimmune myasthenia gravis (EAMG) 192193194195. It also can suppress autoreactivity in vitro to two respective autoantigens, MBP and acetylcholine receptor (AChR). This has lead to the suggestion that it may be one of several factors responsible for inhibition of disease activity during the third trimester of pregnancy in patients with MS as well as MG. It is of some interest that expression of the gene for this potentially immunosuppressive protein is downregulated by Th1 and M/M cytokines.

Nestin, an intermediate filament protein, is a marker of early neuronal cell development. It is also a marker of other progenitor cells, particularly glial cells in the CNS, and may be involved in cell proliferation 196197198199. The proinflammatory Th1 cytokine mixture down regulated the gene for nestin, which would be compatible with an inhibitory effect of such cytokines on neuronal and glial cell precursors.

VIPR2 binds VIP, a peptide shown to induce release of cytokines and other factors from glial cells 200201. Downregulation of this protein by Th1 cytokines secreted by infiltrating inflammatory cells or endogenous glia would inhibit the release of both cytokines and growth factors by glial cells.

We detected M/M induced upregulation of a gene transcript for angiotensin receptor 2 (ATR2), whereas Th1 cytokines down regulated expression of the same transcript and Th2 cytokines down regulated a different ATR2 transcript. M/M cytokines upregulated the gene for ATR 1. ATR 2 is expressed by endothelial cells as well as glial cells. ATR1 is also expressed by endothelial cells as well as other cells within the CNS. Angiotensin and ATR are involved with interactions with VEGF and other molecules and may be involved in CNS cell death via apoptosis 202203204. Increased expression of the gene for angiotensin, the ligand for angiotensin receptors, has been described in studies of MS brain tissue 205206.

Two unexpected and novel findings were the marked decreases in expression of the genes for chapsyn-110 and beta tubulin by Th2 cytokines. Chapsyn-110 is a member of the PSD95/SAP90 protein family. The protein is found in postsynaptic densities in somatic/dendritic neuronal processes, and interacts with the C-terminus of subunits of the NMDA GluR and shaker-type potassium channel 207208. The protein is linked indirectly to microtubules and involved in clustering of the receptors and ion channels; its presence and function in glia have not been previously reported. The marked down regulation of chapsyn-110 along with that of beta tubulin in glia may lead to potentially neuroprotective disruption of signaling through NMDA receptors and potassium channels in these cells.

Ceruloplasmin is a metal binding protein which is increased in response to inflammatory signals. In the brain ceruloplasmin is important as a binder of iron, and in the absence of ceruloplasmin (aceruloplasminemia), iron is able to induce tissue injury by increasing lipid peroxidation 209210211. M/M cytokines upregulated the expression of the gene for ceruloplasmin whereas Th1 and Th2 cytokines had no effect. Effects on genes for iron binding proteins if resulting in sufficient increase in protein would down regulate free iron induced lipid peroxidation, whereas a reduction or even a failure of increase in such proteins could result in cell damage or even death.

Caveolins are a group of proteins that are important in the structure of cell membranes including neurons and myelin. They are components of the so called "lipid rafts", important constituents of plasma membranes. Caveolins 1, 2 and 3 are upregulated in spinal cord of rat with EAE with caveolin 3 being expressed by astrocytes 212, although at 6 hours in vitro Th1 cytokines down regulated the expression of the gene for caveolin 3.

Arginase 1 is involved in synthesis of polyamines which have been shown to improve axonal regeneration on myelin substrates 213. Th2 upregulated the gene for this protein, which would favor axonal regeneration. Th2 cytokines, particularly IL-4, stimulate production of arginase by macrophages, and there is an inverse relationship between production of iNOS induced by Th1 cytokines and arginase induced by Th2 cytokines in these cells 214215216. By inhibiting production of nitric oxide, arginase may also play a neuroprotective role for motor neurons deprived of trophic factors 217. Recently, loss of arginase 1 was shown to increase proliferation of neural stem cells 218. One could postulate that the microglia may be the glial cells upregulating the gene for arginase in our system.

It has been reported that certain MS lesions have features characteristic of ischemic or hypoxic injury to oligodendrocytes 12219 although inflammatory cells, particularly macrophages, are present in the lesions. Studies of normal appearing white matter in MS, employing gene array technology, have also shown changes in patterns of gene regulation consistent with ischemia and the response to ischemia 9. It has also been suggested that local angiogenesis occurs in EAE and in MS 220. We have identified early effects of these cytokine mixtures on molecules that are important in vascular pathology and angiogenesis as well as upregulated in ischemia and hypoxia. Obviously this includes a vast number of genes and gene products involved in transcription, cell signaling, mitochondrial function and apoptosis along with many others.

In addition to the changes in genes for proteins associated with apoptosis and mitochondria, in our current and prior studies 1516, we found that Th1 cytokines upregulated other genes reported to be regulated in the CNS in ischemia including adhesion molecules (ICAM-1, VCAM), cytokines and chemokines and their receptors (IL-1β, MCP-1), death and survival proteins (Bcl-X), proteases and inhibitors (MMP-9) and growth factors (FGF 1 and NGFRp75). Th1 cytokines did not affect the gene for e-selectin but upregulated the gene for its ligand. Among other genes for proteins regulated in CNS ischemia, Th1 cytokines down regulated genes for neurotrophins and their receptors (BDNF, NT3 and trkB), and cytokines, chemokines and their relevant receptors (several related to TGF-β). M/M cytokines upregulated genes for cell adhesion molecules (ICAM-1, VCAM), HSP 70, cytokines, chemokines and receptors (IL-1β, IL-1R type 1, IL-6, MCP-1), FGF 5 and 10, and MMPs and inhibitors (MMP9, TIMP-1) and downregulated genes of interest for response to ischemia, including TGF-β3, NT3, and FGF2. Th2 cytokines upregulated ischemia related genes for growth factors (BDNF, FGF 10 and 14, FGF-R1), cytokines and chemokines and receptors (IL-6, IL-1R type I and TGF-βR2) and down regulated genes for IL-1R type I, and NT3. Differential expression of many of these genes were reported in the NAWM of some patients with MS 9.

As previously reported 1516, Th1, M/M and Th2 cytokines had varying effects on the genes for molecules that are involved in altering in the cells of the blood brain barrier including several adhesion molecules and MMPs although our cultures do not contain endothelial cells. Some of these molecules are undoubtedly important in glial cells as well. In a previous study, we detected upregulation of the gene for VEGF 16. Upregulation of VEGF could contribute to endothelial cell proliferation seen in some MS lesions producing local hypoxia and oligodendroglial death. The function of VEGF in glial cells as well as other non-glial non-neuronal cells, such as pericytes, which conceivably might be in our cultures is not known. Since inflammatory cytokines were able to upregulate the gene for VEGF as well as other genes that are associated with ischemia and the response to ischemia, our data suggests that cytokine release secondary to inflammation can lead to changes compatible with hypoxia and perhaps to induction of hypoxia itself.

We recognize the limitations of microarray analysis as well as gene expression studies since post-transcriptional and post-translational changes are not detected. In addition proteins such as receptors may be present in sufficient amount to be ligated and involved in a biologic process without requiring additional protein in the short run and thus no upregulation of gene for that protein. Nevertheless as a screening technique to obtain an overview of proteins that may be important in a particular process as well as the complexities of the effect of a mixture of factors on a mixture of cells, we believe that this is a promising approach. In addition microarray technology allows discovery of unexpected findings in complex experiments. Such findings may turn out to be both interesting and important.